PRISMA flow diagram examples, with the numbers explained

Follow the count at each stage, then load the same example into the editor. These fictional examples explain the arithmetic; they are not findings from published reviews.

Fictional teaching data. All five examples use the same database-search trail so you can see what changes when another branch or a previous review is added. The examples have no ongoing or awaiting-classification reports.

1. New review · databases

A new systematic review using databases and registers.

Databases and registers: complete count trail
Stage	Count	How it is accounted for
Identified	1,200	PubMed: 700; Embase: 300; Trial registers: 200
Removed before the next stage	250	200 duplicates; 20 by automation; 30 for other reasons
Records screened	950	1,200 − 200 − 20 − 30
Records excluded at screening	850	Title / abstract decisions
Reports sought for retrieval	100	950 screened − 850 excluded; no record-to-report adjustment
Reports not retrieved	10	Not assessed for eligibility
Reports assessed for eligibility	90	100 sought − 10 not retrieved
Excluded after assessment	60	Ineligible population: 40; Ineligible study design: 20
Included reports from this branch	30	90 assessed − 60 excluded

The current search contributes 30 reports linked to 24 studies. Different report and study counts are expected when some studies have more than one report.

PRISMA 2020 flow diagram example for a new systematic review searching databases and registers: 1,200 records identified from databases and registers, 950 records screened, 90 reports assessed for eligibility, 24 studies and 30 reports included. — The completed diagram for these fictional counts. Open the full-size SVG.

Load this example in the editor →

2. New review · all sources

Databases and registers, plus websites, organisations or citation searching.

Databases and registers: complete count trail
Stage	Count	How it is accounted for
Identified	1,200	PubMed: 700; Embase: 300; Trial registers: 200
Removed before the next stage	250	200 duplicates; 20 by automation; 30 for other reasons
Records screened	950	1,200 − 200 − 20 − 30
Records excluded at screening	850	Title / abstract decisions
Reports sought for retrieval	100	950 screened − 850 excluded; no record-to-report adjustment
Reports not retrieved	10	Not assessed for eligibility
Reports assessed for eligibility	90	100 sought − 10 not retrieved
Excluded after assessment	60	Ineligible population: 40; Ineligible study design: 20
Included reports from this branch	30	90 assessed − 60 excluded

Other sources: complete count trail
Stage	Count	How it is accounted for
Identified	60	Websites: 40; Citation searching: 20
Removed before the next stage	10	10 duplicates; 0 by automation; 0 for other reasons
Reports sought for retrieval	50	60 identified − 10 duplicates
Reports not retrieved	5	Not assessed for eligibility
Reports assessed for eligibility	45	50 sought − 5 not retrieved
Excluded after assessment	25	Ineligible population: 15; Ineligible study design: 10
Included reports from this branch	20	45 assessed − 25 excluded

The branches contribute 30 + 20 = 50 included reports. This example has zero remaining overlap between branches; duplicates have already been removed. The linked evidence represents 40 studies.

PRISMA 2020 flow diagram example for a new systematic review that also searched other sources: 1,200 records identified from databases and registers, 950 records screened, 90 reports assessed for eligibility, 60 records from websites and citation searching, 40 studies and 50 reports included. — The completed diagram for these fictional counts. Open the full-size SVG.

Load this example in the editor →

3. Updated review · databases

An update with a separate previous-review arm.

Databases and registers: complete count trail
Stage	Count	How it is accounted for
Identified	1,200	PubMed: 700; Embase: 300; Trial registers: 200
Removed before the next stage	250	200 duplicates; 20 by automation; 30 for other reasons
Records screened	950	1,200 − 200 − 20 − 30
Records excluded at screening	850	Title / abstract decisions
Reports sought for retrieval	100	950 screened − 850 excluded; no record-to-report adjustment
Reports not retrieved	10	Not assessed for eligibility
Reports assessed for eligibility	90	100 sought − 10 not retrieved
Excluded after assessment	60	Ineligible population: 40; Ineligible study design: 20
Included reports from this branch	30	90 assessed − 60 excluded

The current search contributes 30 reports and 24 newly included studies; some current reports also describe retained previous studies. Different report and study counts are expected when some studies have more than one report.

The previous review contained 10 studies and 12 reports. One previous study, represented by two reports, no longer meets eligibility. The update therefore retains 9 studies and 10 reports.

Final study count: 10 − 1 + 24 = 33. Final report count: 12 − 2 + 30 = 40.

Here, 24 is the count of newly included studies. Three of the current 30 reports describe retained previous studies. Those three reports are already a subset of the current report total; do not add them again or count them as three new studies. The study-to-report links need to be checked in the review log.

PRISMA 2020 flow diagram example for an updated systematic review searching databases and registers: 1,200 records identified from databases and registers, 950 records screened, 90 reports assessed for eligibility, 33 studies and 40 reports in the updated review. — The completed diagram for these fictional counts. Open the full-size SVG.

Load this example in the editor →

4. Updated review · all sources

Previous studies, a new database search and other sources.

Databases and registers: complete count trail
Stage	Count	How it is accounted for
Identified	1,200	PubMed: 700; Embase: 300; Trial registers: 200
Removed before the next stage	250	200 duplicates; 20 by automation; 30 for other reasons
Records screened	950	1,200 − 200 − 20 − 30
Records excluded at screening	850	Title / abstract decisions
Reports sought for retrieval	100	950 screened − 850 excluded; no record-to-report adjustment
Reports not retrieved	10	Not assessed for eligibility
Reports assessed for eligibility	90	100 sought − 10 not retrieved
Excluded after assessment	60	Ineligible population: 40; Ineligible study design: 20
Included reports from this branch	30	90 assessed − 60 excluded

Other sources: complete count trail
Stage	Count	How it is accounted for
Identified	60	Websites: 40; Citation searching: 20
Removed before the next stage	10	10 duplicates; 0 by automation; 0 for other reasons
Reports sought for retrieval	50	60 identified − 10 duplicates
Reports not retrieved	5	Not assessed for eligibility
Reports assessed for eligibility	45	50 sought − 5 not retrieved
Excluded after assessment	25	Ineligible population: 15; Ineligible study design: 10
Included reports from this branch	20	45 assessed − 25 excluded

The branches contribute 30 + 20 = 50 included reports. This example has zero remaining overlap between branches; duplicates have already been removed. The current search adds 40 newly included studies; some current reports also describe retained previous studies.

The previous review contained 10 studies and 12 reports. One previous study, represented by two reports, no longer meets eligibility. The update therefore retains 9 studies and 10 reports.

Final study count: 10 − 1 + 40 = 49. Final report count: 12 − 2 + 50 = 60.

Here, 40 is the count of newly included studies. Three of the current 50 reports describe retained previous studies. Those three reports are already a subset of the current report total; do not add them again or count them as three new studies. The study-to-report links need to be checked in the review log.

PRISMA 2020 flow diagram example for an updated systematic review that also searched other sources: 1,200 records identified from databases and registers, 950 records screened, 90 reports assessed for eligibility, 60 records from websites and citation searching, 49 studies and 60 reports in the updated review. — The completed diagram for these fictional counts. Open the full-size SVG.

Load this example in the editor →

5. Scoping review · PRISMA-ScR

The ScR adaptation uses the selection stages to track evidence sources. Read these counts alongside the kinds of evidence your protocol allows.

Databases and registers: complete count trail
Stage	Count	How it is accounted for
Identified	1,200	PubMed: 700; Embase: 300; Trial registers: 200
Removed before the next stage	250	200 duplicates; 20 by automation; 30 for other reasons
Records screened	950	1,200 − 200 − 20 − 30
Records excluded at screening	850	Title / abstract decisions
Sources sought for retrieval	100	950 screened − 850 excluded; no record-to-report adjustment
Sources not retrieved	10	Not assessed for eligibility
Sources assessed for eligibility	90	100 sought − 10 not retrieved
Excluded after assessment	60	Ineligible population: 40; Ineligible study design: 20
Included sources of evidence	30	90 assessed − 60 excluded

Outcome: 30 sources of evidence. These could include different evidence types if eligible under the scoping-review protocol. There is no requirement to turn this count into a meta-analysis total.

PRISMA-ScR flow diagram example for a scoping review: 1,200 records identified from databases and registers, 950 records screened, 90 sources of evidence assessed for eligibility, 30 sources of evidence included. — The completed diagram for these fictional counts. Open the full-size SVG.

Load this example in the editor →

An error example: 10 records too many

After identifying 1,200 records, this log removes 200 duplicates, 20 records by automation and 30 for other reasons. The next stage must account for 950 screened records: 1,200 − 200 − 20 − 30 = 950.

If the diagram says 960, it contains 10 records too many at that stage. Check whether the identified total, removal categories or screened total was copied incorrectly. A balancing change without a supporting log entry hides the problem.

A second frequent mistake is to call 10 unavailable full texts “excluded after assessment.” If 100 reports were sought and 10 were not retrieved, 90 were assessed. Only decisions about those 90 reports belong in the eligibility-exclusion reasons.

Sources and attribution

Journal Metrics provides an independent implementation of the PRISMA flow diagram. The original PRISMA 2020 templates are available under CC BY 4.0. Our editor adapts their layout and adds count checks; it is not endorsed by the PRISMA group.

Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.
Page MJ, Moher D, Bossuyt PM, et al. PRISMA 2020 explanation and elaboration. BMJ. 2021;372:n160.
Official PRISMA 2020 checklist and expanded checklist — use these to review the full manuscript.
Tricco AC, Lillie E, Zarin W, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med. 2018;169:467–473. See the official ScR checklist.

Content updated October 13, 2026. Count examples are illustrative.